Ace the 2026 Anthropic Fellows Challenge: Master AI Safety, Economics & Research Methods!

The idea being tested is how AI can affect work in two different ways: displacement and augmentation, and what that means for policy. When AI-driven displacement occurs, automation takes over tasks that people used to perform, reducing the demand for those tasks and potentially leading to job losses in those areas. Augmentation, on the other hand, increases human productivity by handling routine or data-heavy parts of a job and letting people focus on more complex or creative tasks; this tends to keep or raise the overall demand for human labor.

Policy implications follow those patterns. If displacement is prominent, policies should help people adapt: retraining programs so workers can move into higher-skill roles, unemployment support during transitions, and safety nets to cushion income as workers shift careers. If augmentation is more common, the emphasis shifts to enabling effective collaboration between humans and AI—investing in skill development, education, and systems that help workers leverage AI to boost productivity.

The other options don’t fit: the claim that displacement increases demand or that augmentation reduces tasks contradicts how these effects actually function; treating displacement and augmentation as the same ignores their distinct labor-market impacts; and saying displacement affects only capital misses the human-workforce dimension that policy must address.